70 million Americans suffer from some sort of sleep disorder. How the brain switches between states of consciousness is not known, but researchers have now determined that a distributed network of neurons are responsible for wake, non-REM and REM sleep. First-order circuit models have emerged, and these have been tested using traditional lesion and gene deletion methods. We now propose to repair the network using the gene transfer method. This approach has proven to be an effective neurobiological tool in a number of neurodegenerative diseases but surprisingly no one has used it to correct a sleep disorder. In this project recombinant adeno-associated viral vector (rAAV-serotype 8) will be used to transfer the gene for the arousal peptide hypocretin (also known as orexin) into a specific neuronal phenotype. An integrated series of four aims will test specific hypotheses in two murine models with targeted deletions of the hypocretin gene or of the hypocretin neurons. Experiments with appropriate controls are proposed to strengthen the conclusions. These aims are supported by our very strong preliminary data that shows robust expression of the hypocretin gene product in specific neurons along with unambiguous decline of narcoleptic behavior in the two murine models. These studies will demonstrate that genetic pharmacology can serve as a valuable neurobiological tool to repair the network.